This document describes the known issues and limitations for Gemini Enterprise.
Connecting and managing data
The following are known issues and limitations when connecting and managing your data in Gemini Enterprise:
You can only have 500 data stores per project. If you need more data stores, you need to use a new project.
You can only have 500 Gemini Enterprise apps per project. If you need more apps, you need to use a new project.
After you create an app, the Google Cloud project and location are fixed for the data store and app. Data and settings cannot be automatically migrated between different projects or locations.
You cannot connect your website data to a Gemini Enterprise app.
An app can only support one data connection method (federated or ingestion) for a connector. If the connector supports both data federation and data ingestion, select your preferred data connection method.
Using Gemini 3.1 Pro and 3 Flash in Limited Availability
Gemini Enterprise users now have access to 3.1 Pro and 3 Flash in Limited Availability as a Pre-GA Offering but with soon-to-be announced General Availability SLOs. The goal is to enable our users to use these 3 series models in their production workloads.
The following are specific constraints and known limitations for the deployment of select Gemini 3 models within Gemini Enterprise in Limited Availability:
Deployment status
- Limited Availability: The Gemini 3 series models are currently offered as a Pre-GA Offering - Limited Availability status within Gemini Enterprise. "Limited Availability" as used here means that while this model deployment will provide, in time, production-grade General Availability (GA) SLOs, as a Pre-GA Offering we don't provide full GA contractual commitments such as lifetime support or standardized 12-month deprecation notice for changes to the model or its availability.
- Cloud Data Processing Addendum: As a Pre-GA Offering with Limited Availability status, you may use Gemini 3.1 Pro and 3 Flash as of the last updated date in Gemini Enterprise to process personal data as outlined in the Cloud Data Processing Addendum.
Access limitations
- No Direct API Access: The specific Limited Availability model capabilities are available exclusively through the Gemini interface and integrated surfaces. For this specific deployment, you won't have access to these models using the Vertex AI API.
Regional and data residency limitations
Gemini Enterprise features are subject to specific limitations based on the selected region and API type. See Gemini Enterprise Standard and Plus Editions data residency and ML regional processing commitments for more information.
- DRZ (Data Residency): Means your data at rest (stored) and in transit does not leave a specified geographic boundary (like the EU, US, or a specific country).
- MLP (Machine Learning Processing): Means the actual AI computation, model inference, and data processing happens within the selected geographic region.
The following table outlines the Data Residency commitments for Gemini 3.x models in the Gemini Enterprise app:
| Feature | US/EU/global multiregion | In Country regions (IN, SG, UK, CA) |
|---|---|---|
| Gemini 3.1 Pro | DRZ supported. MLP supported. | DRZ coming soon. MLP coming soon. |
| Gemini 3 Flash | DRZ supported. MLP supported. | DRZ coming soon. MLP coming soon. |
Using Nano Banana Pro in Preview for image generation
The following are known issues and limitations when using Nano Banana Pro (Gemini 3 Pro Image) in Preview with Gemini Enterprise:
Nano Banana Pro is in Preview for image generation and is only available in the Vertex AI Global region. Therefore, it doesn't meet the data residency (DRZ) or machine learning (ML) processing requirements in US or EU regions for Gemini Enterprise.
After an admin enables Nano Banana Pro for image generation in Gemini Enterprise Standard or Plus edition, it might take some time to become visible on the web app.
We recommend that you don't generate more than 3 images per chat request.